Debian Bug report logs - #294180
ifrename should be run before ifplugd

Toggle useless messages

Message #3 received at submit@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: hook scripts for net class

Date: Tue, 8 Feb 2005 12:02:28 +0100

[Message part 1 (text/plain, inline)]

Package: hotplug
Version: 0.0.20040329-16
Severity: wishlist

ifplugd and waproamd both use /etc/hotplug.d/net/* scripts to start
themselves when a new network card becomes available. This bites if
programmes like ifrename or udev want to rename an interface:
ifplugd/waproamd ups the interface before /etc/hotplug.d/default/*
scripts are run. Thus, when udev/ifrename get a chance, the
interface is already up and cannot be renamed.

The solution -- I think -- seems to be to call ifplugd/waproamd from
/etc/hotplug/net/* instead. However, that does not exist.

Could you please add support for hook scripts for the net class to
/etc/hotplug?

Thanks,

-- System Information:
Debian Release: 3.1
  APT prefers testing
  APT policy: (600, 'testing'), (98, 'unstable'), (1, 'experimental')
Architecture: i386 (i686)
Kernel: Linux 2.6.10-wing
Locale: LANG=en_GB, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages hotplug depends on:
ii  debconf                      1.4.42      Debian configuration management sy
ii  grep                         2.5.1.ds1-4 GNU grep, egrep and fgrep
ii  module-init-tools            3.2-pre1-2  tools for managing Linux kernel mo
ii  modutils                     2.4.26-1.2  Linux module utilities
ii  procps                       1:3.2.4-1   The /proc file system utilities

-- debconf information excluded

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #8 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: martin f krafft <madduck@debian.org>, 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Tue, 8 Feb 2005 12:27:56 +0100

[Message part 1 (text/plain, inline)]

On Feb 08, martin f krafft <madduck@debian.org> wrote:

> Could you please add support for hook scripts for the net class to
> /etc/hotplug?
I need to think about this and discuss this with the upstream
maintainer. Don't hold your breath.

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #11 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Marco d'Itri <md@Linux.IT>

Cc: 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Tue, 8 Feb 2005 12:59:14 +0100

[Message part 1 (text/plain, inline)]

also sprach Marco d'Itri <md@Linux.IT> [2005.02.08.1227 +0100]:
> I need to think about this and discuss this with the upstream
> maintainer. Don't hold your breath.

I won't. However, please consider elevating the priority since right
now, ifplugd and waproamd effectively disable ifrename in some
configurations (not enough to conflict), and the same applies for
udev's renaming abilities.

Given that the net class does not really need the usermap
functionality, I think it would suffice to run-parts the
/etc/hotplug/net directory with the appropriate environment. Then
again, maybe it would be better to stick to a common approach and
employ usermaps.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #16 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: martin f krafft <madduck@debian.org>, 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 01:53:11 +0100

[Message part 1 (text/plain, inline)]

On Feb 08, martin f krafft <madduck@debian.org> wrote:

> The solution -- I think -- seems to be to call ifplugd/waproamd from
> /etc/hotplug/net/* instead. However, that does not exist.
How should such a directory work?
Do you feel it is still needed, considering that /etc/hotplug/usb/ will
not be supported by hotplug-ng?

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #19 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Marco d'Itri <md@Linux.IT>

Cc: 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 07:54:24 +0100

[Message part 1 (text/plain, inline)]

also sprach Marco d'Itri <md@Linux.IT> [2005.02.16.0153 +0100]:
> > The solution -- I think -- seems to be to call ifplugd/waproamd
> > from /etc/hotplug/net/* instead. However, that does not exist.
> How should such a directory work?

run-parts with the information in the environment?

> Do you feel it is still needed, considering that /etc/hotplug/usb/
> will not be supported by hotplug-ng?

I do not know too much about hotplug-ng. I am also not sure what
hotplug/usb has to do with my proposal about hotplug/net. We are
talking about hooks to trigger for network devices, whether USB or
not...

Please enlighten me.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #24 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: martin f krafft <madduck@debian.org>

Cc: 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 11:13:16 +0100

[Message part 1 (text/plain, inline)]

On Feb 16, martin f krafft <madduck@debian.org> wrote:

> ... then the only problem that remains is ifplugd and renaming.
> Will hotplug-ng rename the interfaces/device nodes before calling
> the hook scripts, or is the renaming done with the
> /etc/hotplug.d/default scripts again?
hotplug-ng does not renames network interfaces. But if you use udevsend
as the hotplug multiplexer then I think that udev will rename the
interfaces before anything else.

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #27 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Marco d'Itri <md@Linux.IT>

Cc: 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 11:18:16 +0100

[Message part 1 (text/plain, inline)]

also sprach Marco d'Itri <md@Linux.IT> [2005.02.16.1113 +0100]:
> hotplug-ng does not renames network interfaces. But if you use udevsend
> as the hotplug multiplexer then I think that udev will rename the
> interfaces before anything else.

Will Debian do so? Otherwise ifplugd, waproamd clash with this
renaming.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #32 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: martin f krafft <madduck@debian.org>

Cc: 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 11:24:25 +0100

[Message part 1 (text/plain, inline)]

On Feb 16, martin f krafft <madduck@debian.org> wrote:

> also sprach Marco d'Itri <md@Linux.IT> [2005.02.16.1113 +0100]:
> > hotplug-ng does not renames network interfaces. But if you use udevsend
> > as the hotplug multiplexer then I think that udev will rename the
> > interfaces before anything else.
> Will Debian do so? Otherwise ifplugd, waproamd clash with this
> renaming.
Yes, starting with the next hotplug package. You can already try this by
upgrading to the latest udev package and modifying the hotplug init
script to stop changing /proc/sys/kernel/hotplug.

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #35 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Marco d'Itri <md@Linux.IT>

Cc: 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 11:36:01 +0100

[Message part 1 (text/plain, inline)]

also sprach Marco d'Itri <md@Linux.IT> [2005.02.16.1124 +0100]:
> Yes, starting with the next hotplug package. You can already try this by
> upgrading to the latest udev package and modifying the hotplug init
> script to stop changing /proc/sys/kernel/hotplug.

Awesome. Is this going to be in sarge?

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #38 received at 294180-submitter@bugs.debian.org (full text, mbox, reply):

From: jdthood@aglu.demon.nl

To: md@Linux.IT (Marco d'Itri), 294180-submitter@bugs.debian.org

Subject: Re: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 11:27:21 GMT

> Any comments?

This is a long-standing problem.  I know of at least the following
reports that relate to this:

    #245435: ifplugd.hotplug: Please ifrename the i'face first
    #267007: waproamd: Needs ifrename support since its hotplug
             script is called before net.agent
    #294111: ifplugd: iflugd breaks udev when customizing names
             of network devices
    #294180: hook scripts for net class

Originally I filed #245435 against hotplug with the title "ifrename
shouldn't be in net.agent" and asked that hotplug move the ifrename
call to /sbin/hotplug so that it would happen before hook scripts
were called.

Since the hotplug maintainers were unresponsive at the time and the
ifrename maintainer adopted the position that ifplugd and waproamd
hook scripts shouldn't expect the interface already to have been
renamed when they run (since this is outside "the normal flow"), I
thought it expedient to reassign the report to ifplugd to see if
its maintainer was willing to do anything about the problem.  If
the ifplugd hook script were to call ifrename itself then the
problem would be solved, although it would result in ifrename
getting called more than once (an otherwise harmless waste of CPU
cycles).

But no.  The ifplugd maintainer turned out to be another zombie.
After a few months I tagged my report "wontfix" since I thought
that the problem really should really be fixed at the hotplug level
anyway.  I saw signs on linux-hotplug-devel that the problem had
been recognized and would be solved.  However, it seems that that
is not the case.  (I haven't followed the issue since then because
I solved the problem locally by adding an ifrename call to
/sbin/hotplug and none of the involved maintainers seemed to care
about it.)

I thank Md for looking into the problem.

How to solve the problem?  Running ifrename from /sbin/hotplug
instead of from net.agent works for me.  I would send a patch but
I am not at my Debian machine right now.  Anyway, other solutions
can be imagined.  Instead of running ifrename earlier you can run
the ifplugd and waproamd hook scripts later, as martin krafft
suggested.  I don't have any opinion about what the best
approach is because I haven't been reading linux-hotplug-devel
and consequently don't know what the development plans are for
hotplug and udev.  I could sit down and study how hotplug and
udev interact today and come up with some solution which works
both when hotplug is installed without udev, and when when it is
installed with udev, and calls ifrename before the ifplugd and
waproamd hook scripts run, and doesn't call ifrename more than
once for any given "add" event.  Also, my preferred solution
would put the actual ifrename call into a hook script owned by
the ifrename package, so that the hotplug maintainers don't have
to worry about it again.  But I think that Md can just as easily
and more competently do this since he knows where hotplug and
udev are going in the future.

-- 
Thomas

Message #43 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: jdthood@aglu.demon.nl

Cc: 294180@bugs.debian.org

Subject: Re: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 12:52:21 +0100

[Message part 1 (text/plain, inline)]

On Feb 16, jdthood@aglu.demon.nl wrote:

> How to solve the problem?  Running ifrename from /sbin/hotplug
> instead of from net.agent works for me.  I would send a patch but
This will not happen, hotplug-ng uses a C program and anyway on udev
systems udevsend will be used as the hotplug multiplexer.
OTOH, when udevsend is used then udev should be able to rename the
interfaces before other hotplug scripts are run.

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #46 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: jdthood@aglu.demon.nl, 294180-quiet@bugs.debian.org

Cc: Marco d'Itri <md@Linux.IT>

Subject: Re: Bug#294180: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 13:07:39 +0100

[Message part 1 (text/plain, inline)]

I can confirm that the latest hotplug (not yet uploaded) and the
latest udev (not yet uploaded) fix this issue halfway. Now udevsend
is used for the notifications, and so udev can rename the iface
before ifplugd and waproamd get called. So now the problem is
ifrename. Given that /etc/hotplug is going to go anyway, I think
ifrename should be fixed to run before ifplugd, e.g. as
/etc/hotplug.d/net/000ifrename.hotplug.

Thomas: I would be happy to hijack ifplugd, waproamd, and ifrename,
given that the maintainers are zombies. Would you be co-maintainer?

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #51 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: martin f krafft <madduck@debian.org>

Cc: jdthood@aglu.demon.nl, 294180@bugs.debian.org

Subject: Re: Bug#294180: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 13:17:01 +0100

[Message part 1 (text/plain, inline)]

reassign 294180 ifrename
retitle 294180 ifrename should be run before ifplugd
thanks

On Feb 16, martin f krafft <madduck@debian.org> wrote:

> I can confirm that the latest hotplug (not yet uploaded) and the
> latest udev (not yet uploaded) fix this issue halfway. Now udevsend
> is used for the notifications, and so udev can rename the iface
> before ifplugd and waproamd get called. So now the problem is
> ifrename. Given that /etc/hotplug is going to go anyway, I think
> ifrename should be fixed to run before ifplugd, e.g. as
> /etc/hotplug.d/net/000ifrename.hotplug.
Agreed, I'm reassigning the bug. (Maybe 000 is excessive, let's start
with something like 020_ifrename.)

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #60 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: jdthood@aglu.demon.nl

To: martin f krafft <madduck@debian.org>, jdthood@aglu.demon.nl, 294180-quiet@bugs.debian.org, Marco d'Itri <md@Linux.IT>

Subject: Re: Bug#294180: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 12:26:14 GMT

> I can confirm that the latest hotplug (not yet uploaded) and the
> latest udev (not yet uploaded) fix this issue halfway. Now udevsend
> is used for the notifications, and so udev can rename the iface
> before ifplugd and waproamd get called. So now the problem is
> ifrename. Given that /etc/hotplug is going to go anyway, I think
> ifrename should be fixed to run before ifplugd, e.g. as
> /etc/hotplug.d/net/000ifrename.hotplug.


There remains the problem that the wrong interface name is provided
to the ifplugd hook script in the environment variable.


> Thomas: I would be happy to hijack ifplugd, waproamd, and ifrename,
> given that the maintainers are zombies. Would you be co-maintainer?


ifrename is part of wireless-tools and is, so far as I know, well
maintained.

waproamd has been orphaned upstream and should probably be removed
from Debian.  I raised this issue on debian-release
    http://lists.debian.org/debian-release/2005/01/msg00138.html
but got no response.

Some months ago the ifplugd maintainer agreed to let me take over
maintainership of ifplugd.  I repackaged it using cdbs.  When the
maintainer saw that I had thrown away his work he decided that he
didn't want to let the package go after all.  So he held onto it
and went back to sleep.  In Debian, there is nothing you can do
about a stubborn zombie unless he becomes completely MIA.  I don't
know how Oliver Kurth will react if you approach him.  Perhaps he
has learned something over the past months.
-- 
Thomas Hood

Message #65 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: martin f krafft <madduck@debian.org>

Cc: jdthood@aglu.demon.nl, 294180@bugs.debian.org

Subject: Re: Bug#294180: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 13:39:03 +0100

[Message part 1 (text/plain, inline)]

On Feb 16, martin f krafft <madduck@debian.org> wrote:

> Marco: if udevsend replaces /sbin/hotplug, then udev should be
> installed on the standard system, huh? Maybe then we can either
> integrate ifrename into it, or argue that ifrename is not necessary
> anymore?
ifrename is not needed if you are using udev. If you are not, you are on
your own.

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #68 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: jdthood@aglu.demon.nl

Cc: 294180-quiet@bugs.debian.org, Marco d'Itri <md@Linux.IT>

Subject: Re: Bug#294180: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 13:36:35 +0100

[Message part 1 (text/plain, inline)]

also sprach jdthood@aglu.demon.nl <jdthood@aglu.demon.nl> [2005.02.16.1326 +0100]:
> Some months ago the ifplugd maintainer agreed to let me take over
> maintainership of ifplugd.  I repackaged it using cdbs.  When the
> maintainer saw that I had thrown away his work he decided that he
> didn't want to let the package go after all.  So he held onto it
> and went back to sleep.  In Debian, there is nothing you can do
> about a stubborn zombie unless he becomes completely MIA.  I don't
> know how Oliver Kurth will react if you approach him.  Perhaps he
> has learned something over the past months.

I was his sponsor and advocate. I will talk to him.

About waproamd: could you simply file a request to remove the
package against ftp.debian.org? Maybe we can make udev and ifrename
conflict?

About ifrename, good to know it's maintained.

Marco: if udevsend replaces /sbin/hotplug, then udev should be
installed on the standard system, huh? Maybe then we can either
integrate ifrename into it, or argue that ifrename is not necessary
anymore?

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #71 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Marco d'Itri <md@Linux.IT>

Cc: jdthood@aglu.demon.nl, 294180-quiet@bugs.debian.org

Subject: Re: Bug#294180: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 13:50:48 +0100

[Message part 1 (text/plain, inline)]

also sprach Marco d'Itri <md@Linux.IT> [2005.02.16.1339 +0100]:
> ifrename is not needed if you are using udev. If you are not, you are on
> your own.

Sure. I wonder whether in the future, ifrename makes any sense.

  - for it to work, it must be run at the right time, which is in
    /etc/init.d/networking, or a hotplug scripts.

  - in the future, /etc/init.d/networking will go away, in favour of
    hotplugging (i think, though there are still ISA cards and the
    like).

  - hotplug is soon not the hotplug provider anymore as udevsend
    takes over.

  - udevsend comes with udev, meaning that udev has to be installed.

  - thus, ifrename is not needed anymore.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #76 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: jdthood@aglu.demon.nl

To: martin f krafft <madduck@debian.org>, jdthood@aglu.demon.nl, 294180-quiet@bugs.debian.org, Marco d'Itri <md@Linux.IT>

Subject: Re: Bug#294180: [madduck@debian.org: Bug#294180: hook scripts for net class]

Date: Wed, 16 Feb 2005 13:14:25 GMT

> About waproamd: could you simply file a request to remove the
> package against ftp.debian.org? Maybe we can make udev and ifrename
> conflict?


Not being the waproamd author or maintainer or a release manager
(or even a DD) and not having found any support on the lists, I
don't feel justified in doing that.


> Marco: if udevsend replaces /sbin/hotplug, then udev should be
> installed on the standard system, huh? Maybe then we can either
> integrate ifrename into it, or argue that ifrename is not necessary
> anymore?


udev makes ifrename unnecessary.  However, at present udev requires
Linux 2.6 and Linux 2.6 isn't the only kernel that Debian supports.
So, ifrenaming of hot plugged network interfaces should still be
implemented correctly.  It isn't too hard to do.

-- 
Thomas
Hood

Message #81 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Jean Tourrilhes <jt@hpl.hp.com>

To: 294180@bugs.debian.org, md@linux.it, jdthood@aglu.demon.nl

Subject: Re: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 13:50:49 -0800

Marco wrote :
> 
> ifrename is not needed if you are using udev. If you are not, you are on
> your own.

	Sorry, but that's not true. It pains me that the hotplug/udev
project suffers from NIH, this is not the first time I see them making
blanket statements that their stuff completely replace the old stuff
when it's not (yet) the case.

	udev doesn't understand much about networking (that is not its
job) and therefore is lacking critical functionality when it comes to
renaming network interfaces. Just one simple example : udev can't
distinguish between the two network interfaces created by the Aironet
driver for each card, and therefore will give them the same name.
	Yes, udev is nice and work for 95% of the cases. For the
remaining 5%, you really need ifrename. As ifrename handles 100% of
the cases, I personally would not waste any time with udev for network
interfaces, and rather would work at better integrating ifrename in
the udev/hotplug framework. Note that by the time you fix udev to
handle all the subtelty of network interfaces, you will have the
perfect clone of ifrename ;-)

	Have fun...

	Jean

Message #86 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Jean Tourrilhes <jt@hpl.hp.com>

To: 294180@bugs.debian.org, jdthood@aglu.demon.nl

Subject: Re: Bug#294180: hook scripts for net class

Date: Wed, 16 Feb 2005 14:04:27 -0800

martin f krafft wrote :
> 
> About waproamd: could you simply file a request to remove the
> package against ftp.debian.org? Maybe we can make udev and ifrename
> conflict?

	That's counter productive. The main goal of udev is to
populate /dev, and the interface name renaming is just a side
feature. There are many scenario where you would want to use udev and
ifrename together.

Thomas wrote :
> udev makes ifrename unnecessary.

	For someone familiar with ifrename and its documentation, I
find this statement puzzling. Care to explain how you reached this
conclusion ?
	Thanks...

	Jean

Message #91 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Thomas Hood <jdthood@aglu.demon.nl>

To: jt@hpl.hp.com

Cc: 294180@bugs.debian.org

Subject: Re: Bug#294180: hook scripts for net class

Date: Thu, 17 Feb 2005 08:19:46 +0100

On Wed, 2005-02-16 at 14:04 -0800, Jean Tourrilhes wrote:
> Thomas wrote :
> > udev makes ifrename unnecessary.
> 
> 	For someone familiar with ifrename and its documentation, I
> find this statement puzzling.


OK, I change my statement to "udev makes ifrename unnecessary in many
cases."

I agree with you that ifrename should be supported.  It should certainly
be supported when hotplug is installed and udev is not.  I would like to
see it supported when both hotplug and udev are installed but the udev
maintainer may not want that.

-- 
Thomas Hood <jdthood@aglu.demon.nl>

Message #96 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Guus Sliepen <guus@sliepen.eu.org>

To: Thomas Hood <jdthood@aglu.demon.nl>, 294180@bugs.debian.org

Cc: jt@hpl.hp.com, md@Linux.IT, martin f krafft <madduck@debian.org>, Fumitoshi UKAI <ukai@debian.or.jp>, Oliver Kurth <oku@debian.org>

Subject: Interface renaming

Date: Thu, 17 Feb 2005 15:17:13 +0100

[Message part 1 (text/plain, inline)]

severity 294180 important
merge 294180 291111
thanks

On Thu, Feb 17, 2005 at 08:19:46AM +0100, Thomas Hood wrote:

> OK, I change my statement to "udev makes ifrename unnecessary in many
> cases."
> 
> I agree with you that ifrename should be supported.  It should certainly
> be supported when hotplug is installed and udev is not.  I would like to
> see it supported when both hotplug and udev are installed but the udev
> maintainer may not want that.

I've followed the discussion and I understand the current situation is
problematic, mainly because there are so many ways to detect and bring
up network interfaces. Although I understand why both #294180 and
#291111 have been (re)assigned to ifrename, there is not much I can do
about it (call me a zombie if you want). The ifrename package does not
contain an init script (I don't think anyone knows which priority it
should have) nor hook scripts for any other package. 

Rather than reassigning bugs again, I have created a Wiki page where we
can discuss this problem in a more constructive way (hopefully):

http://wiki.debian.net/?InterfaceRenaming

Please add your comments there.

-- 
Met vriendelijke groet / with kind regards,
    Guus Sliepen <guus@sliepen.eu.org>

[signature.asc (application/pgp-signature, inline)]

Message #103 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Guus Sliepen <guus@sliepen.eu.org>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, 294180@bugs.debian.org, jt@hpl.hp.com, md@Linux.IT, Fumitoshi UKAI <ukai@debian.or.jp>, Oliver Kurth <oku@debian.org>

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 17:18:56 +0100

[Message part 1 (text/plain, inline)]

also sprach Guus Sliepen <guus@sliepen.eu.org> [2005.02.17.1517 +0100]:
> Rather than reassigning bugs again, I have created a Wiki page where we
> can discuss this problem in a more constructive way (hopefully):
> 
> http://wiki.debian.net/?InterfaceRenaming

I do not find a Wiki to be a discussion tool.

Can't we stick to email, create a mailing list for this, or move to
debian-devel?

> Please add your comments there.

Done.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #108 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Guus Sliepen <guus@sliepen.eu.org>

To: martin f krafft <madduck@debian.org>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, 294180@bugs.debian.org, jt@hpl.hp.com, md@Linux.IT, Fumitoshi UKAI <ukai@debian.or.jp>, Oliver Kurth <oku@debian.org>

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 17:42:09 +0100

[Message part 1 (text/plain, inline)]

On Thu, Feb 17, 2005 at 05:18:56PM +0100, martin f krafft wrote:

> > http://wiki.debian.net/?InterfaceRenaming
> 
> I do not find a Wiki to be a discussion tool.
> 
> Can't we stick to email, create a mailing list for this, or move to
> debian-devel?

Well, everything is better than having the discussion spread over
several bug reports. At least the Wiki can provide a summary of the
issues. As this hopefully won't be a long-lasting discussion, I don't
think creating a mailing list would be worth it, so... debian-devel then?

> > Please add your comments there.
> 
> Done.

Thanks!

-- 
Met vriendelijke groet / with kind regards,
    Guus Sliepen <guus@sliepen.eu.org>

[signature.asc (application/pgp-signature, inline)]

Message #111 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Guus Sliepen <guus@sliepen.eu.org>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, 294180@bugs.debian.org, jt@hpl.hp.com, md@Linux.IT, Fumitoshi UKAI <ukai@debian.or.jp>, Oliver Kurth <oku@debian.org>

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 17:48:28 +0100

[Message part 1 (text/plain, inline)]

also sprach Guus Sliepen <guus@sliepen.eu.org> [2005.02.17.1742 +0100]:
> Well, everything is better than having the discussion spread over
> several bug reports. At least the Wiki can provide a summary of the
> issues. As this hopefully won't be a long-lasting discussion, I don't
> think creating a mailing list would be worth it, so... debian-devel then?

Can't we just merge all bug reports for now? Or agree on one?

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #116 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Thomas Hood <jdthood@aglu.demon.nl>

To: Guus Sliepen <guus@sliepen.eu.org>

Cc: 294180@bugs.debian.org, jt@hpl.hp.com, md@Linux.IT, martin f krafft <madduck@debian.org>

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 21:11:57 +0100

On Thu, 2005-02-17 at 15:17 +0100, Guus Sliepen wrote:
> The ifrename package does not
> contain an init script (I don't think anyone knows which priority it
> should have)


You can create one at S:S40ifrename.

This was requested and discussed in #243382.  In that thread, J.T. said
at one point that the ifrename call should come after the setting up of
ipv4 kernel parameters, which happens in S:S40networking.  However, I
later discussed this with Md and he convinced me that there was no
reason to insist on this sequence of events.  It should be fine to do
ifrename just before S:S40networking.


>  nor hook scripts for any other package. 


Adding a hotplug hook script is complicated by the fact that the name of
the interface is passed to /etc/hotplug.d/net/*.hotplug scripts in the
INTERFACE environment variable.  It won't do for ifrename to install a
hook script at /etc/hotplug.d/net/000ifrename.hotplug and for
ifplugd.hotplug and waproamd.hotplug to get called with the original
interface name in INTERFACE.

Either
* hotplug should run ifrename and call ifplugd.hotplug and
  waproamd.hotplug with INTERFACE set to the new name
or else
* ifplugd.hotplug and waproamd.hotplug should each call ifrename
  for the interface named in INTERFACE and then should start
  ifplugd/waproamd with the new interface name.

-- 
Thomas Hood <jdthood@aglu.demon.nl>

Message #119 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Thomas Hood <jdthood@aglu.demon.nl>

Cc: Guus Sliepen <guus@sliepen.eu.org>, 294180@bugs.debian.org, jt@hpl.hp.com, md@Linux.IT

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 21:26:26 +0100

[Message part 1 (text/plain, inline)]

also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.02.17.2111 +0100]:
> * hotplug should run ifrename and call ifplugd.hotplug and
>   waproamd.hotplug with INTERFACE set to the new name
> or else
> * ifplugd.hotplug and waproamd.hotplug should each call ifrename
>   for the interface named in INTERFACE and then should start
>   ifplugd/waproamd with the new interface name.

Clearly, then, the first should be favoured.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #124 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Thomas Hood <jdthood@aglu.demon.nl>

To: martin f krafft <madduck@debian.org>

Cc: Guus Sliepen <guus@sliepen.eu.org>, 294180@bugs.debian.org, jt@hpl.hp.com, md@Linux.IT

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 21:25:42 +0100

On Thu, 2005-02-17 at 21:26 +0100, martin f krafft wrote:
> also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.02.17.2111 +0100]:
> > * hotplug should run ifrename and call ifplugd.hotplug and
> >   waproamd.hotplug with INTERFACE set to the new name
> > or else
> > * ifplugd.hotplug and waproamd.hotplug should each call ifrename
> >   for the interface named in INTERFACE and then should start
> >   ifplugd/waproamd with the new interface name.
> 
> Clearly, then, the first should be favoured.


Why do you favor the first over the second?

-- 
Thomas Hood <jdthood@aglu.demon.nl>

Message #127 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Thomas Hood <jdthood@aglu.demon.nl>

Cc: Guus Sliepen <guus@sliepen.eu.org>, 294180@bugs.debian.org, jt@hpl.hp.com, md@Linux.IT

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 21:37:23 +0100

[Message part 1 (text/plain, inline)]

also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.02.17.2125 +0100]:
> > also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.02.17.2111 +0100]:
> > > * hotplug should run ifrename and call ifplugd.hotplug and
> > >   waproamd.hotplug with INTERFACE set to the new name
> > > or else
> > > * ifplugd.hotplug and waproamd.hotplug should each call ifrename
> > >   for the interface named in INTERFACE and then should start
> > >   ifplugd/waproamd with the new interface name.
> > 
> > Clearly, then, the first should be favoured.
> 
> Why do you favor the first over the second?

Sorry, should be more explicit. I favour the first over the second
because it handles interface renaming at the same location as udev
does (approximately). Then ifplugd, waproamd, foobar, and xyz do not
ever need to worry about it. If ifrename's syntax changes, only
hotplug needs to be changed.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #132 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Jean Tourrilhes <jt@hpl.hp.com>

To: martin f krafft <madduck@debian.org>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, Guus Sliepen <guus@sliepen.eu.org>, 294180@bugs.debian.org, md@linux.it

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 12:45:09 -0800

On Thu, Feb 17, 2005 at 09:37:23PM +0100, martin f krafft wrote:
> 
> If ifrename's syntax changes, only
> hotplug needs to be changed.

	The only syntax changes that I could imagine is changes to
make it easier to integrate with hotplug (this was actually the reason
I rewrote nameif which did not integrate in hotplug). If you have
suggestions, let's talk about it.
	One think on my mind is what's an efficient way to get the
path to sysfs (I know how to do it the slow way). Either pass it on
the command line or in an environment variable.
	Have fun...

	Jean

Message #135 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Jean Tourrilhes <jt@hpl.hp.com>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, Guus Sliepen <guus@sliepen.eu.org>, 294180@bugs.debian.org, md@linux.it

Subject: Re: Interface renaming

Date: Thu, 17 Feb 2005 21:49:51 +0100

[Message part 1 (text/plain, inline)]

also sprach Jean Tourrilhes <jt@hpl.hp.com> [2005.02.17.2145 +0100]:
> 	The only syntax changes that I could imagine is changes to
> make it easier to integrate with hotplug (this was actually the reason
> I rewrote nameif which did not integrate in hotplug). If you have
> suggestions, let's talk about it.

Sure, though Marco is the better person here.

I was simply making a point; whether ifrename changes syntax or not
does not matter. The renaming should be handled as early as possible
so that other packages do not have to know about it.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #140 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Thomas Hood <jdthood@aglu.demon.nl>

To: 294180@bugs.debian.org, 294180-submitter@bugs.debian.org

Cc: Jean Tourrilhes <jt@hpl.hp.com>, md@linux.it

Subject: Re: Interface renaming

Date: Fri, 18 Feb 2005 09:42:35 +0100

> also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.02.17.2125 +0100]:
> > > also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.02.17.2111 +0100]:
> > > > * hotplug should run ifrename and call ifplugd.hotplug and
> > > >   waproamd.hotplug with INTERFACE set to the new name
> > > > or else
> > > > * ifplugd.hotplug and waproamd.hotplug should each call ifrename
> > > >   for the interface named in INTERFACE and then should start
> > > >   ifplugd/waproamd with the new interface name.
> > > 
> > > Clearly, then, the first should be favoured.
> > 
> > Why do you favor the first over the second?
> 
> Sorry, should be more explicit. I favour the first over the second
> because it handles interface renaming at the same location as udev
> does (approximately). Then ifplugd, waproamd, foobar, and xyz do not
> ever need to worry about it. If ifrename's syntax changes, only
> hotplug needs to be changed.


The first option may have advantages but Md does not seem to favor it.


On Feb 16, I wrote:
> How to solve the problem?  Running ifrename from /sbin/hotplug
> instead of from net.agent works for me.

and Md replied:
> This will not happen, hotplug-ng uses a C program and anyway on udev
> systems udevsend will be used as the hotplug multiplexer.
> OTOH, when udevsend is used then udev should be able to rename the
> interfaces before other hotplug scripts are run.


(It is possible that Md's views have changed since he wrote this, of
course.  This is cc:ed to him so he can tell us what he thinks now.)

There is another option, actually (call it 'option #3') which is for
ifrename to wrap /sbin/hotplug with a script that runs ifrename, changes
INTERFACE and execs the original /sbin/hotplug.

  * Divert /sbin/hotplug to /sbin/hotplug.hotplug
  * Include /sbin/hotplug which ifrenames and execs /sbin/hotplug.hotplug

This should work both with hotplug and with the future hotplug-ng.  This
wouldn't handle the case where the future udev is installed that does:

  echo /sbin/udevsend > /proc/sys/kernel/hotplug

If we want ifrename to work with that future udev then /sbin/udevsend
will have to be wrapped in a similar way (... if we adopt option #3).

-- 
Thomas Hood <jdthood@aglu.demon.nl>

Message #146 received at 294180@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Thomas Hood <jdthood@aglu.demon.nl>, 294180-quiet@bugs.debian.org

Cc: 294180@bugs.debian.org, 294180-submitter@bugs.debian.org, Jean Tourrilhes <jt@hpl.hp.com>, md@linux.it

Subject: Re: Bug#294180: Interface renaming

Date: Fri, 18 Feb 2005 10:25:16 +0100

[Message part 1 (text/plain, inline)]

also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.02.18.0942 +0100]:
> There is another option, actually (call it 'option #3') which is
> for ifrename to wrap /sbin/hotplug with a script that runs
> ifrename, changes INTERFACE and execs the original /sbin/hotplug.
[...]
> If we want ifrename to work with that future udev then
> /sbin/udevsend will have to be wrapped in a similar way (... if we
> adopt option #3).

I was going to say... if ifrename is being used and udev installed,
ifrename must not just stop working.

Thus, the solution you suggest is a viable one.

Whether this is done with diversions, or by actually writing
/sbin/ifrename.hotplug to /proc/sys/kernel/hotplug and then checking in the
/sbin/ifrename.hotplug script whether /sbin/udevsend is available to
call, and only fall back to /sbin/hotplug otherwise... we'll have to
discuss more I guess. I favour the second approach (not with
diversions) because it seems cleaner.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #151 received at 294180@bugs.debian.org (full text, mbox, reply):

From: md@Linux.IT (Marco d'Itri)

To: Thomas Hood <jdthood@aglu.demon.nl>

Cc: 294180@bugs.debian.org, Jean Tourrilhes <jt@hpl.hp.com>

Subject: Re: Interface renaming

Date: Fri, 18 Feb 2005 10:29:22 +0100

[Message part 1 (text/plain, inline)]

On Feb 18, Thomas Hood <jdthood@aglu.demon.nl> wrote:

> > This will not happen, hotplug-ng uses a C program and anyway on udev
> > systems udevsend will be used as the hotplug multiplexer.
> > OTOH, when udevsend is used then udev should be able to rename the
> > interfaces before other hotplug scripts are run.
> (It is possible that Md's views have changed since he wrote this, of
> course.  This is cc:ed to him so he can tell us what he thinks now.)
No changes.

> There is another option, actually (call it 'option #3') which is for
> ifrename to wrap /sbin/hotplug with a script that runs ifrename, changes
> INTERFACE and execs the original /sbin/hotplug.
Probably not an option, if all goes well /sbin/hotplug will be replaced
with a daemon listening for uevents on a PF_NETLINK socket.

-- 
ciao,
Marco

[signature.asc (application/pgp-signature, inline)]

Message #162 received at 294180@bugs.debian.org (full text, mbox, reply):

From: jdthood@aglu.demon.nl

To: 294180@bugs.debian.org, 294180-submitter@bugs.debian.org

Cc: md@Linux.IT (Marco d'Itri), Jean Tourrilhes <jt@hpl.hp.com>

Subject: Re: Interface renaming

Date: Fri, 18 Feb 2005 10:48:06 GMT

severity 295514 wishlist
severity 294111 wishlist
severity 267007 wishlist
severity 295520 wishlist
tags 245435 - wontfix
tags 267007 - wontfix
merge 294111 295514 245435
merge 267007 295520 
thanks


> Probably not an option, if all goes well /sbin/hotplug will be replaced
> with a daemon listening for uevents on a PF_NETLINK socket.

OK, strike options #1 and #3.  That leaves #2.

-- 
Thomas

Message #170 received at 294180@bugs.debian.org (full text, mbox, reply):

From: jdthood@aglu.demon.nl

To: 294180@bugs.debian.org, 294180-submitter@bugs.debian.org

Cc: Marco d'Itri <md@Linux.IT>

Subject: Re: Bug#294180: Interface renaming

Date: Fri, 18 Feb 2005 11:37:15 GMT

martin f krafft wrote:
> Can a hook in /etc/hotplug.d stop /sbin/hotplug from executing the
> rest of the scripts? In that case, ifrename could just call
> /sbin/hotplug again, or submit a uevent to the PF_NETLINK socket,
> causing hotplug to call all hook again.


Call this 'option #4'.

Too ugly, IMHO.  I hate it when things re-run themselves.


> As an alternative, we will have to provide a little tool that
> ifplugd and waproamd can call (or better, source a file) to handle
> the interface renaming in one single location.


Sounds like overkill, given that we are only talking about two packages
maintained by the same person, one of which will disappear before too
long.

-- 
Thomas
Hood

Message #176 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: jdthood@aglu.demon.nl, 294180-quiet@bugs.debian.org

Cc: Marco d'Itri <md@Linux.IT>, Jean Tourrilhes <jt@hpl.hp.com>

Subject: Re: Bug#294180: Interface renaming

Date: Fri, 18 Feb 2005 17:50:48 +0100

[Message part 1 (text/plain, inline)]

also sprach jdthood@aglu.demon.nl <jdthood@aglu.demon.nl> [2005.02.18.1237 +0100]:
> Call this 'option #4'.
> 
> Too ugly, IMHO.  I hate it when things re-run themselves.

udev does it.

> > As an alternative, we will have to provide a little tool that
> > ifplugd and waproamd can call (or better, source a file) to handle
> > the interface renaming in one single location.
> 
> Sounds like overkill, given that we are only talking about two packages
> maintained by the same person, one of which will disappear before too
> long.

who knows how many other packages will be affected in the future?

and remember: if you find yourself doing something twice, you did
something wrong the first time.

aside, semantics of how ifrename should be checked for and called
should be within the domain of ifrename; ifplugd should not need to
posess any knowledge.

also sprach jdthood@aglu.demon.nl <jdthood@aglu.demon.nl> [2005.02.18.1240 +0100]:
> What do you mean by "the result was the exact opposite"?

they ignored my suggestion for improvement of a feature and instead
removed the whole feature, without clear reasons. check the
linux-hotplug-devel archives for posts about permissions.d. there is
also a udev bug.

also sprach jdthood@aglu.demon.nl <jdthood@aglu.demon.nl> [2005.02.18.1344 +0100]:
> If Md isn't against it then I'd say that we should try to get
> upstream to implement #1 before we fall back to #2.

please try, and i wish you more luck than i had.

maybe refrain from pointing them at the bug(s)? not sure...

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #179 received at 294180-submitter@bugs.debian.org (full text, mbox, reply):

From: Thomas Hood <jdthood@aglu.demon.nl>

To: 294180-submitter@bugs.debian.org

Subject: NMU?

Date: Sun, 13 Mar 2005 13:24:57 +0100

> also sprach jdthood@aglu.demon.nl <jdthood@aglu.demon.nl> [2005.02.18.1344 +0100]:
> > If Md isn't against it then I'd say that we should try to get
> > upstream to implement #1 before we fall back to #2.
> 
> please try, and i wish you more luck than i had.


I haven't done anything about this, have you?

I say we NMU ifplugd and waproamd to add the necessary ifrename calls
to /etc/hotplug.d/net/{ifplugd,waproamd}.hotplug.  The result is some
superfluous ifrename calls but at least things will work properly.

-- 
Thomas Hood <jdthood@aglu.demon.nl>

Message #182 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Thomas Hood <jdthood@aglu.demon.nl>, 294180-quiet@bugs.debian.org

Cc: oku@debian.org

Subject: Re: Bug#294180: NMU?

Date: Sun, 13 Mar 2005 23:40:05 +0100

[Message part 1 (text/plain, inline)]

also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.03.13.1324 +0100]:
> I haven't done anything about this, have you?

Well, I spoke to Oliver, who said (IIRC) that he was going to look
at it. I guess he has not found the time.

I think an NMU is appropriate. Do you have any objections, Oliver?

Thomas, I am really loaded right now, so I would appreciate if you
could handle this.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #187 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: Oliver Kurth <okurth@gmx.net>

To: martin f krafft <madduck@debian.org>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, 294180-quiet@bugs.debian.org, oku@debian.org

Subject: Re: Bug#294180: NMU?

Date: Sun, 13 Mar 2005 22:27:05 -0800

[Message part 1 (text/plain, inline)]

On Sun, 2005-03-13 at 23:40 +0100, martin f krafft wrote:
> also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.03.13.1324 +0100]:
> > I haven't done anything about this, have you?
> 
> Well, I spoke to Oliver, who said (IIRC) that he was going to look
> at it. I guess he has not found the time.
> 
> I think an NMU is appropriate. Do you have any objections, Oliver?
> 
> Thomas, I am really loaded right now, so I would appreciate if you
> could handle this.

I think I would prefer a well tested patch, and I will upload it myself.
There already is another NMU coming (debconf), and two NMUs would
conflict. I promise to take some time for this tomorrow afternoon
(Monday, Pacific time).

I am really not happy about this. IMHO ifrename is an ugly hack. Ugly
hacks need other ugly hacks. I would prefer the hotplug solution if any,
but it seems that you had no success with that.

Greetings,
Oliver

[signature.asc (application/pgp-signature, inline)]

Message #192 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: Oliver Kurth <okurth@gmx.net>

To: martin f krafft <madduck@debian.org>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, 294180-quiet@bugs.debian.org, oku@debian.org

Subject: Re: Bug#294180: NMU?

Date: Mon, 14 Mar 2005 14:50:16 -0800

[Message part 1 (text/plain, inline)]

On Sun, 2005-03-13 at 22:27 -0800, Oliver Kurth wrote:
> On Sun, 2005-03-13 at 23:40 +0100, martin f krafft wrote:
> > also sprach Thomas Hood <jdthood@aglu.demon.nl> [2005.03.13.1324 +0100]:
> > > I haven't done anything about this, have you?
> > 
> > Well, I spoke to Oliver, who said (IIRC) that he was going to look
> > at it. I guess he has not found the time.
> > 
> > I think an NMU is appropriate. Do you have any objections, Oliver?
> > 
> > Thomas, I am really loaded right now, so I would appreciate if you
> > could handle this.
> 
> I think I would prefer a well tested patch, and I will upload it myself.
> There already is another NMU coming (debconf), and two NMUs would
> conflict. I promise to take some time for this tomorrow afternoon
> (Monday, Pacific time).
> 
> I am really not happy about this. IMHO ifrename is an ugly hack. Ugly
> hacks need other ugly hacks. I would prefer the hotplug solution if any,
> but it seems that you had no success with that.

Okay, I spent some time with ifrename, and here is my solution for the
hotplug script. Please test it, if it works for you I will upload a new
version soon.

Greetings,
Oliver

[ifplugd.hotplug (application/x-shellscript, attachment)]

[signature.asc (application/pgp-signature, inline)]

Message #197 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Jean Tourrilhes <jt@hpl.hp.com>

To: 294180@bugs.debian.org, jdthood@aglu.demon.nl, oku@debian.org, madduck@cirrus.madduck.net

Subject: [PATCH] My own viewpoint on this

Date: Mon, 14 Mar 2005 18:23:04 -0800

	Hi,

	This bug has gone around enough, it's time to find a solution
and fix it for good.

	This is my solution :

-----------------------------------------------------------
diff -u -p ifplugd-0.26/debian/hotplug-old ifplugd-0.26/debian/hotplug
--- ifplugd-0.26/debian/hotplug-old	2005-03-14 18:02:18.000000000 -0800
+++ ifplugd-0.26/debian/hotplug	2005-03-14 18:05:21.000000000 -0800
@@ -32,6 +32,18 @@ case $ACTION in
 add|register)
 	for IF in $HOTPLUG_INTERFACES ; do
 		if [ "$INTERFACE" = "$IF" -o "$IF" = "all" ] ; then
+			# Run ifrename as needed - Jean II
+			# This is required because we bypass net.agent
+			# where it's normally done.
+			if [ -x /sbin/ifrename ] && [ -r /etc/iftab ]; then
+				debug_mesg invoke ifrename for $INTERFACE
+				NEWNAME=`/sbin/ifrename -i $INTERFACE`
+				if [ -n "$NEWNAME" ]; then
+					debug_mesg iface $INTERFACE is remapped to $NEWNAME
+					INTERFACE=$NEWNAME
+				fi;
+			fi
+
 			debug_mesg Invoking $DAEMON_NAME for $INTERFACE
 			/etc/init.d/$DAEMON_NAME start $INTERFACE
 			break
-----------------------------------------------------------

	My comments :

	1) ifrename was made to be called from various places, so I
don't mind adding one more. It's called from hotplug, but also from
the standard network init scripts. It cal also be called from
/etc/network/interfaces if you want. That's why the command line has
so many options.

	2) I personally feel that running ifplugd from hotplug is
completely wrong for various reasons. I strongly believe that it
should be called from ifup.
		a) Not all interface configuration goes through
hotplug. I may want to enable interfaces manually.
		b) I want to control which interface are ifplugd or
not in /etc/network/interfaces, where all my configuration reside. I
don't want network configuration scatered in zillion places.
		c) ifplugd is a super-dhcp, as far as I am concerned,
I don't know why we should treat it differently than dhcp (which is
handled by ifup).
		d) Not deterministic. If I add another daemon in
hotplug that grab and configure interfaces, how does the system chose
if ifplugd should have it first or the other daemon ? Or, maybe the
scripts will change and run net.agent before ifplugd.
	I agree that this goal is hard to acheive in the short term,
knowing the state of ifup, and hotplug is an acceptable short term
solution. But, now that Thomas is on the case, I think we can make
progress.

	Have fun...

	Jean

Message #202 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Oliver Kurth <okurth@gmx.net>

To: jt@hpl.hp.com

Cc: 294180@bugs.debian.org, jdthood@aglu.demon.nl, oku@debian.org, madduck@cirrus.madduck.net

Subject: Re: [PATCH] My own viewpoint on this

Date: Mon, 14 Mar 2005 21:50:56 -0800

[Message part 1 (text/plain, inline)]

On Mon, 2005-03-14 at 18:23 -0800, Jean Tourrilhes wrote:
> 	Hi,
> 
> 	This bug has gone around enough, it's time to find a solution
> and fix it for good.
> 
> 	This is my solution :

Thanks. Mine is very similar, except that I call ifrename a little
earliar, so that it is possible to enable/disable the renamed interface.
I sent the script with my last mail to this bug, a few hours ago.

Greetings,
Oliver

[signature.asc (application/pgp-signature, inline)]

Message #207 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Thomas Hood <jdthood@aglu.demon.nl>

To: jt@hpl.hp.com

Cc: 294180@bugs.debian.org, oku@debian.org, madduck@cirrus.madduck.net

Subject: Re: [PATCH] My own viewpoint on this

Date: Tue, 15 Mar 2005 08:21:43 +0100

On Mon, 2005-03-14 at 18:23 -0800, Jean Tourrilhes wrote:
> 	This is my solution :
[...]

On Mon, 2005-03-14 at 21:50 -0800, Oliver Kurth wrote: 
> Thanks. Mine is very similar, except that I call ifrename a little
> earliar, so that it is possible to enable/disable the renamed interface.
> I sent the script with my last mail to this bug, a few hours ago.


Yes, O.K.'s solution looks better.


On Mon, 2005-03-14 at 18:23 -0800, Jean Tourrilhes wrote:
> 	2) I personally feel that running ifplugd from hotplug is
> completely wrong for various reasons. I strongly believe that it
> should be called from ifup.


I don't fully agree with that, but I don't think it is worth disputing
the point because I am sure that we both agree that the current network
configuration system is fundamentally inadequate.  ifupdown was designed
to control statically configured interfaces.  Something more intelligent
is needed in order to deal with the dynamic networking of today.

-- 
Thomas Hood <jdthood@aglu.demon.nl>

Message #210 received at 294180-quiet@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Oliver Kurth <okurth@gmx.net>

Cc: Thomas Hood <jdthood@aglu.demon.nl>, 294180-quiet@bugs.debian.org, oku@debian.org

Subject: Re: Bug#294180: NMU?

Date: Tue, 15 Mar 2005 09:08:11 +0100

[Message part 1 (text/plain, inline)]

also sprach Oliver Kurth <okurth@gmx.net> [2005.03.14.2350 +0100]:
> Okay, I spent some time with ifrename, and here is my solution for
> the hotplug script. Please test it, if it works for you I will
> upload a new version soon.

Looks good. I cannot test it at the moment, but it should do exactly
what we need.

-- 
 .''`.     martin f. krafft <madduck@debian.org>
: :'  :    proud Debian developer, admin, user, and author
`. `'`
  `-  Debian - when you have better things to do than fixing a system
 
Invalid/expired PGP subkeys? Use subkeys.pgp.net as keyserver!

[signature.asc (application/pgp-signature, inline)]

Message #217 received at 294180@bugs.debian.org (full text, mbox, reply):

From: Jean Tourrilhes <jt@hpl.hp.com>

To: Thomas Hood <jdthood@aglu.demon.nl>

Cc: 294180@bugs.debian.org, oku@debian.org, madduck@cirrus.madduck.net

Subject: Re: [PATCH] My own viewpoint on this

Date: Tue, 15 Mar 2005 09:53:51 -0800

On Tue, Mar 15, 2005 at 08:21:43AM +0100, Thomas Hood wrote:
> On Mon, 2005-03-14 at 18:23 -0800, Jean Tourrilhes wrote:
> > 	This is my solution :
> [...]
> 
> On Mon, 2005-03-14 at 21:50 -0800, Oliver Kurth wrote: 
> > Thanks. Mine is very similar, except that I call ifrename a little
> > earliar, so that it is possible to enable/disable the renamed interface.
> > I sent the script with my last mail to this bug, a few hours ago.
> 
> 
> Yes, O.K.'s solution looks better.

	Sorry for the duplicata ;-)

> On Mon, 2005-03-14 at 18:23 -0800, Jean Tourrilhes wrote:
> > 	2) I personally feel that running ifplugd from hotplug is
> > completely wrong for various reasons. I strongly believe that it
> > should be called from ifup.
> 
> 
> I don't fully agree with that,

	At least, I provide valid reasons for that. The points I list
will need to be addressed somehow.

> but I don't think it is worth disputing
> the point because I am sure that we both agree that the current network
> configuration system is fundamentally inadequate.  ifupdown was designed
> to control statically configured interfaces.  Something more intelligent
> is needed in order to deal with the dynamic networking of today.

	I fully agree. It pains me to see all the ugly workaround in
various packages (especially hotplug/net.agent) for the limitations of
ifupdown.
	On the other hand, we want to retain flexibility and control
in the system. Not every interfaces should be automatically managed by
something like ifplugd, I personally want some of my interfaces to
have static IP and wireless config (with manual schemes). I hate when
the system tries to be more clever than me and doesn't give me
override.

> Thomas Hood <jdthood@aglu.demon.nl>

	Have fun...

	Jean

Message #222 received at 294180-close@bugs.debian.org (full text, mbox, reply):

From: Oliver Kurth <oku@debian.org>

To: 294180-close@bugs.debian.org

Subject: Bug#294180: fixed in ifplugd 0.26-2

Date: Wed, 16 Mar 2005 01:32:16 -0500

Source: ifplugd
Source-Version: 0.26-2

We believe that the bug you reported is fixed in the latest version of
ifplugd, which is due to be installed in the Debian FTP archive:

ifplugd_0.26-2.diff.gz
  to pool/main/i/ifplugd/ifplugd_0.26-2.diff.gz
ifplugd_0.26-2.dsc
  to pool/main/i/ifplugd/ifplugd_0.26-2.dsc
ifplugd_0.26-2_i386.deb
  to pool/main/i/ifplugd/ifplugd_0.26-2_i386.deb



A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 294180@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Oliver Kurth <oku@debian.org> (supplier of updated ifplugd package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Format: 1.7
Date: Mon, 14 Mar 2005 13:33:13 -0800
Source: ifplugd
Binary: ifplugd
Architecture: source i386
Version: 0.26-2
Distribution: unstable
Urgency: low
Maintainer: Oliver Kurth <oku@debian.org>
Changed-By: Oliver Kurth <oku@debian.org>
Description: 
 ifplugd    - A configuration daemon for ethernet devices
Closes: 236856 245435 251860 287513 291111 294180 295514 298589 299428
Changes: 
 ifplugd (0.26-2) unstable; urgency=low
 .
   * clarified /etc/default/ifplugd docs (closes: #287513). Thanks to
     Marc Haber.
   * remove list of allowed ifaces in hotplug script (closes: #251860,
     #295514, #298589)
   * call ifrename in hotplug script (closes: #294180, #291111, #245435)
   * po-debconf (closes: #236856). Thanks to Lucas Wall for support.
   * added pt_BR debconf templates (closes: #299428). Thanks to
     Felipe Augusto van de Wiel.
   * added po files for cs, da, es, fr, zh_TW. Thanks to the submitters.
Files: 
 71abed9b7e48cfb10b146291fa8ee8b2 609 net optional ifplugd_0.26-2.dsc
 c73388d28f228d06e974e24b4f4ddb64 38806 net optional ifplugd_0.26-2.diff.gz
 afed995ce2720e6183d003f22be39068 55928 net optional ifplugd_0.26-2_i386.deb
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQFCN9DCUmVSJkUeqxsRAsAMAJ0WiMD/Pe1XGUjZ5xTsn2jikFmKVQCgkKZI
XT+16JSRZrX6tLpt+JE4evg=
=zqz5
-----END PGP SIGNATURE-----

Send a report that this bug log contains spam.